Skip to main content

Full text of "Psychometric properties of a magnitude matching task using pressure algometry"

See other formats









Unlimited thanks must go to Henry Gremillion, D.D.S., Melvin B. Benson, D.D.S., 
and the staff of the Parker E. Mahan Facial Pain Center at the University of Florida for 
their guidance, support, and supplies. I would also like to thank James McSolay and Gary 
Myers for their time and assistance in the administration of iontophoresis. Finally, love 
and thanks go to my husband, Andrew Cagle, for data collection and, more importantly, 
his unfaltering patience, support, and belief in me. 








Temporomandibular Disorders 1 

Pain Perception and Assessment 2 

Magnitude Matching 4 

Assessment of Pain in TMD 7 

Pressure Algometry 9 

Iontophoresis 12 

Design and Statement of Hypotheses 17 


Subjects 19 

Procedures 19 

Measures 21 



Test-Retest Study 29 

Validity Study 30 

Visual Analog Scales 35 










Abstract of Dissertation Presented to the Graduate School 
of the University of Florida in Partial Fulfillment of the 
Requirements for the Degree of Doctor of Philosophy 



Felicia Brown 

May 1998 

Chairman: Michael E. Robinson, Ph.D. 

Major Department: Clinical and Health Psychology 

Commonly used methods of measuring the subjective experience of pain, such as 
category rating scales, pain threshold measures, cross-modality matching, and magnitude 
estimation, are subject to inherent limitations and biases. Magnitude matching paradigms, 
which require subjects to rate the magnitude of two alternating sensory modalities, an 
experimental and a control stimulus, using a common scale of magnitude for both, provide 
significant improvement over traditional measures. A standardization procedure, whereby 
the experimental stimuli are standardized in terms of the control stimuli, is conducted to 
cancel out individual factors in assigning pain ratings. Thus, magnitude matching is an 
ideal method for examining discriminability and response bias, two components of pain 
experience, and for examining group differences in pain perception. 

The purpose of this research was to examine, in a nonpatient sample, the test- 
retest reliability and validity of a magnitude matching procedure using a pressure 

algometer, a clinically relevant pain stimulus, the importance of which has been 
demonstrated. In studies in the literature on assessment of pain in temporomandibular 
disorders (TMD), the pressure algometer has been found to be a reliable means of 
measuring pressure-pain threshold and a sensitive measure of treatment-response in the 

Test-retest reliability over two occasions, 3-8 days apart, was found to be 
moderate for discriminability (r = .71, p_ < .01) but poor for response bias (r = .44). The 
validity study used iontophoresis as an anesthetic in a double-blind, placebo and no- 
treatment controlled design. Although it was hypothesized that subjects in the anesthesia 
group would demonstrate reduced discriminability as compared with the control groups, 
no differences were found among the 3 groups on this measure. However, differences in 
response bias were found, with both the placebo control and no-treatment control groups 
differing from the experimental group but not from each other. Possible explanations and 
the implications of these findings are discussed. 



Temporomandibular Disorders 

Temporomandibular disorders (TMD) is a general classification for problems that 
are arthrogenous, involving internal derangements within the temporomandibular joint, or 
myogenous, involving extracapsular problems such as muscle and pain dysfunction or 
myofascial pain dysfunction. Symptoms may include pain or tenderness in the region of 
the masticatory muscles, clicking or popping sounds during condylar movements, and 
limitation or deviation of the jaw opening. In an examination of the concerns of 157 
patients seeking treatment for TMD, pain in the masticatory muscles and/or 
temporomandibular joint was the primary complaint in 54% of the sample (Al-Hassan, 
Ismail & Ash, 1986). 

A literature review by Speculand and Goss (1985) attested that up to 88% of the 
general population may have symptoms or signs of TMD at some time, with up to 25% of 
the population experiencing severe symptoms. Only a small proportion of sufferers 
present for treatment, and most patients have no radiographically detectable pathology of 
the TMJ. In addition, the vast majority of persons who present for treatment in clinical 
settings are women. It has been suggested that TMD affects women up to 9 times more 
frequently than men (Green & Marbach, 1982). 

The etiology of TMD is generally believed to be multifaceted and multi-etiological 
(Dworkin, 1995). Psychological factors have been implicated in the etiology, 
pathogenesis, and maintenance of TMD (Brown, Robinson, Riley & Gremillion, 1996). 

Pain Perception and Assessment 

The experience of pain is a subjective phenomenon and cannot be measured 
independently from self-report (Heft & Parker, 1984). For the clinician treating pain 
patients or the researcher studying pain, however, quantification of pain is necessary. The 
methodologies for such quantification have come from psychophysics, which has always 
been concerned with the perception and judgment of sensation magnitude, or intensity—a 
quality that arguably pertains to all sensory modalities (Marks, Szczesiul & Ohlott, 1986). 
It is impossible to discuss pain perception without also discussing measurement issues. 

Several rating scales have typically been employed to obtain some quantitative 
measurement of the experience of pain. Inherently, the sensitivity of such measures is 
limited by the type of rating scale. For instance, category rating scales require the subject 
to choose a number representing the magnitude of pain or a word from a list describing 
the range of pain experience. These scales are restrictive in accuracy to the extent that the 
subject's pain is not exactly described by the word choices. Some category scales are 
graded, allowing the description of pain severity. Heft & Parker (1984), however, have 
shown that the assumption that the categories are equally spaced, on which such scales 
rest, is not valid. Both the position and the semantic meaning of the words in the list 
contribute to the individual's choice of descriptors; the words do not divide the continuum 
of perception into equal portions. 

Improvement over category scales may be obtained by the use of cross-modality 
matching paradigms, which were developed out of the hypothesis that, because intensity is 
a universal attribute of perception, it should be possible to compare the correspondences 
of one sensory mode to another, parallel sensory mode (Marks et al., 1986). A subject is 
thus asked to adjust the intensity of one stimulus modality, such as loudness of a tone, to 
match the perceived intensity of another stimulus modality, such as the brightness of a 
light. The visual analog scale (VAS) is an example of cross-modality matching. A 
common form of this scale provides the subject with a line, often 10 cm long, with the 
instructions to mark the point which illustrates the intensity of his or her pain, given that 
the bottom of the scale represents no pain, whereas the top represents the worst pain 
imaginable. This method allows an infinite number of choice points. However, subjects 
may find them difficult to use because no guidance is provided beyond the endpoints. In 
addition, these scales are subject to regression bias. If a subject indicates that a stimulus 
intensity is near the endpoint of the scale, but a later stimulus would rate even higher in 
magnitude, the subject has no way to indicate this. Cross-modality matching is also not 
well suited for use with some types of stimuli to which the subject may quickly adapt, such 
as thermal sensations (Duncan, Feine, Bushnell & Boyer, 1988) or for stimuli for which 
adjustment is difficult, such as taste or smell (Stevens & Marks, 1980). 

A variant of cross-modal matching, magnitude estimation, solves some of the 
above problems. Here, the subject gives a number that indicates the perceived intensity of 
the stimulus. An unlimited range of responses is then available, eliminating floor and 
ceiling effects. But at the same time, variability is introduced which limits the sensitivity in 
capturing differences between small groups (Duncan et al., 1988). Some studies have also 

shown that some subjects are significantly affected by context effects, such as the range or 
mean intensity of the experimental stimuli, in making cross-modal matches, representing 
an additional source of variability. 

Magnitude Matching 

Magnitude matching, introduced by Stevens and Marks (1980), was developed out 
of the need for a method to measure and compare individual and group variations, 
especially sensory deficits, in the perception of intensity of various stimuli (Marks et al., 
1986). This method incorporates principles from both cross-modality matching and 
magnitude estimation. In these paradigms, the subject is asked to rate the magnitude of 
two alternating sensory modalities, using a common scale of magnitude estimation for 
both. By assigning the same numerical rating to different sensory events, the subject 
indicates that they share the same intensity. Ratings from one modality, the experimental 
stimuli, are then normalized in terms of the other, the control stimuli. Although the 
paradigm assumes that cross-modal matches are absolute, the fact that context may play 
an equal role suggests that the method may be used to assess intergroup differences, as 
long as care is taken when making quantitative assessments; context effects may mitigate 
any differences in discriminability, thereby leading to underestimation of group differences 
(Marks et al., 1986; Stevens & Marks, 1980). 

The normalization procedure, whereby the experimental stimuli are standardized in 
terms of the control stimuli (by dividing mean ratings of the experimental stimuli by the 
grand mean rating of the control stimuli), is conducted to cancel out individual 
idiosyncracies. Examples of these include susceptibility to context effects, range of 
numbers assigned to stimuli, and the absolute size of numbers assigned to stimuli, as well 

as the expectations, past experiences and exposures each subject brings to the 
experimental setting. In theory, these individual factors should be reflected equally across 
modalities of magnitude estimates. Both between-group and subject variability across 
testing sessions is reduced, allowing a subject to serve as his or her own control. The 
result is more consistent data and improved test-retest reliability. Magnitude matching is 
applicable to all continua of stimuli and adaptation is not a problem because only brief 
exposure to fixed levels of stimuli is needed. Moreover, fewer trials are needed, which is 
especially helpful in pain studies, where it is paramount to limit the number of painful 
stimuli to which the subjects are exposed (Duncan et al., 1988; Marks, 1991; Stevens & 
Marks, 1980). 

Duncan and colleagues (1988) were the first to apply a magnitude matching 
paradigm to examine group differences in pain perception. Their cleverly designed study 
involved a simulation of differences in pain perception using as experimental stimuli two 
ranges (45-49 and 46-50 degrees Celsius) of thermal stimuli for normal subjects rather 
than two groups believed to show inherent differences in pain perception. Visual 
brightness was the control stimuli and both magnitude estimation and magnitude matching 
procedures were employed. When the median pain intensity ratings were simply 
compared between the two groups, magnitude estimation did not reveal group differences; 
nor did magnitude matching reveal more than a trend towards higher ratings among 
subjects in the higher temperature group. With the normalizing procedure, however, 
where all pain intensity ratings for each subject were divided by the subject's overall mean 
estimate of control stimuli (visual brightness), robust group differences were found when 
the data was collapsed across the entire range of temperatures and at each of the 

individual temperature comparisons. These authors concluded that subjects are easily able 
to match intensities of pain to another type of stimuli, with the normalization procedure 
increasing sensitivity. Additionally, they suggest that the use of the magnitude matching 
paradigm is most appropriate when making between-group comparisons of pain 
perception or testing the effectiveness of analgesia. 

In making group comparisons of pain perception, it is important to consider both 
discriminability, which is the extent of sensory differences, and response bias, a measure of 
willingness to report pain. Because both absolute and relational components contribute to 
subjects' responses on magnitude matching tasks, it is possible to examine group 
differences while separating out discriminability and response bias. To examine group 
differences, the standardized ratings may be plotted on a graph with stimulus level on the 
X-axis, and the standardized ratings on the Y-axis. When the data are collapsed for all 
subjects in each group, the groups can be represented on the graph with a line identified by 
a slope and intercept, where the former represents discriminability and the latter represents 
response bias. Steeper slopes would indicate a greater change in standardized rating for 
each unit of change in stimulus intensity. Similarly, higher Y-intercepts represent a 
tendency to rate stimuli with a higher magnitude across the range of stimuli. 

Fuller and Robinson (1995) used a magnitude matching paradigm to examine 
differences in pain perception between chronic low back pain patients and nonpatient 
controls. The study employed a clinically relevant pain stimulus, a lumbar extension 
exercise, to compare perception of pain and heaviness, a nonpain stimulus, in these 
subjects. The results of the study indicated that the chronic pain patients were able to 
discriminate between painful levels of stimuli better than the control subjects. However, 

the CLBP patients and controls were equal in discriminability between levels of heaviness. 
In addition, the CLBP patients tended to underestimate the heaviness of weights they were 
lifting, possibly suggesting a tendency to underreport relevant but nonpainful judgments. 
This study highlighted the importance of testing for group differences using clinically 
relevant stimuli rather than experimental pain stimuli. The former seems to be more 
sensitive to features of clinical pain such as duration, affective associations to injury, or 
fear of injury, factors which may not be operative when an experimental pain stimulus is 

Assessment of Pain in TMD 
Tenderness in the masticatory muscles is one of the most common clinical findings 
in TMD (Gracely & Reid, 1995; List, Helkimo & Karlsson, 1991; Ohrbach & Gale, 
1989b). Such tenderness indicates "trigger points," focal areas of muscle tenderness 
within tight bands of the masseter and temporal muscles. No specific histopathological, 
biochemical or electrophysiologic findings are associated with these hyperalgesic spots, 
which may refer pain, tenderness, and autonomic changes, such as redness, swelling and 
sweating to distant locations (Jaeger & Reeves, 1986; McMillan & Blasberg, 1994; 
Reeves, Jaeger & Graff-Radford, 1986) but it has been suggested that CNS processing is 
altered as a consequence of peripheral tissue injury, causing increased sensitivity to both 
painful and innocuous stimuli (Reid, Gracely & Dubner, 1994). Trigger point sensitivity is 
used in the diagnosis of myofascial pain as well as in quantification of the patient's 
experience of pain. The common mode of assessment is manual palpation; firm pressure 
on active trigger points can elicit or intensify spontaneously referred symptoms, whereas 

latent trigger points are tender to palpation but do not elicit referred phenomena (Jaeger & 
Reeves, 1986). 

However, the method of palpation used varies widely, by site and by examiner, 
despite attempts at standardization. A study examining the reliability of clinical findings in 
temporomandibular disorders revealed kappa values ranging from .16 to .45, which are 
moderate at best, for interexaminer reliability of assessment of pain based on palpation of 
each of the masticatory muscles and joints. Summary scores proved somewhat more 
reliable, with kappa values of .51 for pain on muscle palpation and .33 for joint palpation 
(deWijer, Lobbezoo-Scholte, Steenks & Bosman, 1995). Similarly, Dworkin, LeResche 
and DeRouen (1988) examined the reliability of clinical measurements in TMD and 
determined that higher interrater reliability resulted from composite scores of each muscle 
group. In addition, reliability of manual palpation of the masticatory muscles could be 
improved with extensive training on the part of the examiner, but still to only marginal 
levels of reliability. Both interrater and test-retest reliability of manual palpation was 
examined by Stockstill, Gross & McCall (1989). These researchers reported a fair degree 
of interrater reliability that was maintained over a five- week period. 

Reasons for the difficulties in establishing reliability include variations in method of 
palpation, such as which finger tip is used to palpate, the surface area of the finger tip, the 
amount and variation of pressure applied, and the angle of pressure (List, Helkimo & Falk, 
1989). Ohrbach and Gale (1989a) note that the exam procedure may reflect the 
examiner's biases; the force of palpation may be dependent on the perceived severity of 
the undiagnosed condition. In addition, how much verbal input the patient is asked to 
provide, whether an ordinal scale or a visual analog scale is incorporated, and cues 

provided to the examiner by observation of the patient's behavior, such as facial 
movements, can contribute to measurement variability (Gracely & Reid, 1995). Patient 
factors such as variability in signs and symptoms, the subjectivity of pain, and fluctuation 
of symptoms over time also contribute to unreliable results (de Wijer et al., 1995; Gracely 
& Reid, 1995). Variability in method has obvious clinical implications. Additionally, the 
research literature is fraught with vague descriptions of palpation which are not replicable, 
or descriptions allowing replication but lacking established validity (Ohrbach & Gale, 

Pressure Algometry 
A pressure algometer is a hand-held instrument which applies pressure over a 
specific area at a constant, uniform rate, thereby allowing standardization and improving 
the evaluation of muscle tenderness once a trigger point has been located (Gracely & Reid, 
1995). Furthermore, examination with the algometer can replicate the individual's clinical 
pain, suggesting that pressure is an appropriate stimulus to activate substantially similar 
pain processes to those responsible for clinical pain (Ohrbach & Gale, 1989a). In other 
words, the pressure algometer is a clinically relevant pain stimulus, the importance of 
which has been previously pointed out by Fuller & Robinson (1995). 

In the considerable body of literature, the pressure algometer has most commonly 
been used to obtain the pressure-pain threshold (PPT), the subjective point at which a 
gradually increasing pressure sensation becomes painful. Another specific advantage of 
the pressure algometer over manual palpation is that tenderness may be quantified on a 
ratio scale, versus the ordinal scale (e.g., comparatively more or less tender, or mild, 
moderate or severe) rating of tenderness possible with palpation (List et al., 1989). 

Numerous studies have examined the validity and reliability of the pressure algometer for 
obtaining the PPT. 

To demonstrate validity for PPT measures as an index of tenderness, several 
researchers have compared patients with TMD or other types of head and neck pain with 
asymptomatic controls. Seemingly unfailingly, the results have shown that PPTs are lower 
in patient groups, indicating more tenderness among patients (List et al., 1989; McMillan 
&Blasberg, 1994; Ohrbach & Gale, 1989a; Reeves et al., 1986; Reid et al., 1994). 

Many studies have also examined various parameters of reliability of measurement 
of the PPT. Intra- and intersession test-retest reliability have been demonstrated in both 
groups of healthy volunteers (Jensen, Anderson, Olesen and Lindblom, 1986; Ohrbach & 
Gale, 1989b), and TMD patients (List et al., 1989; Reid et al., 1994). Some between- 
session fluctuation has been noted by some authors, perhaps due to learning or a decrease 
in anxiety (Jensen et al., 1986) or, among patients, variation in pain level over time (Reid 
etal., 1994). 

Interrater reliability of PPT measurement has also been assessed in the literature. 
For example, Reeves et al. (1986) found that 2 examiners could reliably arrive at similar 
measurements taken from several different points on the head and neck of myofascial pain 
patients. Moreover, these authors reported a high degree of reliability between 2 
experimenters when locating unmarked trigger points and measuring their sensitivity. 

Other studies attest to the usefulness of the pressure algometer and PPT 
measurement for measuring treatment outcome. First, among head and neck pain patients, 
highly reliable increases in pain-threshold values were demonstrated after passive 
stretching. In nonpatient samples, subcutaneous injection of local anesthesia and 


application of TENS electrical stimulation have both been shown to increase PPTs (Jensen 
et al., 1986; McMillan & Blasberg, 1994; Graff-Radford, Reeves, Baker & Chiu, 1989). 
A study by Bushnell and colleagues (1991) demonstrated that TENS significantly altered 
pain threshold and significantly decreased subjects' ability to detect small differences in 
heat stimuli, whereas placebo TENS altered neither pain threshold nor discriminability 

The results of another study (Graff-Radford, Reeves, Baker & Chiu, 1989) 
indicated that a reduction in pain intensity, as measured by VAS, could occur without a 
concomitant reduction in trigger point sensitivity as a result of TENS treatment. In some 
experimental groups, a significant difference in VAS rating was seen; however, no group 
differences were found in pre- to post-treatment algometer change scores. The authors 
offer possible explanations: the 10 minute application time may not have been sufficient to 
result in pain inhibition, TENS alone is not suitable for treatment of myofascial pain, or the 
sample size may not have been large enough to detect a significant effect. However, the 
authors do not consider the possibility that the experimental manipulation may have had an 
effect on response bias but not on the sensory-discriminative dimension of pain experience, 
thus yielding differences in the VAS measure but not in trigger point sensitivity. 

Together, these treatment outcome studies suggest that pressure algometry is, at 
least to some extent, responsive to intervention in the laboratory. However, because the 
PPT was used to measure sensitivity, it is not certain whether the reductions demonstrated 
reflect a decrease in response bias and/or a decrease in the sensory-discriminative aspects 
of pain. Because it is a pain threshold measure, the PPT is subject to the inherent 
limitations of threshold measurement: subjectivity to placebo effects, response bias, and 

the influence of instructions given, in addition to potential insensitivity to explicit pain- 
reducing manipulations (Price, 1988). 

Some of these shortcomings were avoided in a study in which the authors 
(Svensson, Arendt-Nielsen, Nielsen, & Larsen, 1995) obtained stimulus-response curves in 
addition to the PPT of healthy controls and patients with facial pain. Curves were 
obtained by asking subjects to make cross-modal matches of pain intensity onto visual 
analog scales. Although the results of the study did not demonstrate significantly different 
PPTs between the pain patients and controls, significantly steeper slopes in the S-R curves 
were found in the patient sample versus controls. In addition, injection of local anesthetic 
into the masseter muscles of control subjects significantly increased PPTs and reduced the 
S-R curve slopes relative to baseline measurements, whereas injection of saline increased 
the slopes. PPTs were unaffected by the saline. Although the authors do not discuss the 
issue of response bias versus sensory-discriminative aspects of pain report, it seems 
reasonable to hypothesize that the changes seen in slope do indeed indicate changes in 
sensory-discriminative processes. The PPTs, in contrast, may not have been sensitive 
enough to experimental manipulation to result in change. Svensson et al. note that using 
S-R curves may provide the basis for a triangulation procedure; thus, a magnitude 
matching procedure as we are proposing seems like a logical next-step in the literature. 


Iontophoresis is "the introduction, by means of an electric current, of ions of 
soluble salts into the tissues of the body for therapeutic purposes" (Lark & Gangarosa, 
1990, p. 109). In a recent literature review, Gangarosa, Ozawa, Onkido, Shimomura and 

Hill (1995) conclude that iontophoresis provides an optimal method for application of 
drugs in the treatment of many surface tissues. 

Use of iontophoresis has been in and out of favor since its first applications, which 
may be traced back to the mid 1 700s. Early equipment was crude and did not yield 
reliable results. Despite this, unbridled and overly optimistic claims for the benefits of 
iontophoresis were made, without the delineation of procedures to ensure effectiveness 
and safety (Gangarosa, 1981a; Gangarosa, 1982). Currently, iontophoresis is an "old 
process that has received new life" with the advent of newer technology and drugs and the 
advancement of scientific knowledge regarding transdermal medication delivery 
(Gangarosa, 1988, p. 402). 

In human skin, the primary barrier to percutaneous absorption of any material is 
the stratum corneum, the composition of which is approximately 20% lipids, 40% 
proteins, and 40% water. Electrically charged electrodes will repel an ion that is similarly 
charged; thus ions with a positive charge may be introduced into the body with a positive 
electrode (anode), and negatively ions will be repelled in with a negative electrode 
(cathode) (Byl, Zellerbach & Pfalzer, 1996). Iontophoresis alters skin permeability by 
causing changes in the arrangement of these molecules. According to the "flip-flop gating 
model," when an electric potential, the driving force in iontophoresis, is applied across the 
stratum corneum, certain polypeptide molecules reorient into a parallel arrangement and 
repel neighboring dipoles, forming pores. Hair follicles, sweat gland ducts and sebaceous 
glands act as diffusion shunts, as the reduction in resistance provides pathways for the 
entering molecules. 


For iontophoresis to be clinically effective, the target tissue must be at an 
accessible depth in the body. It does not have to be at the skin surface in order for the 
drug to penetrate in sufficient quantity. Bursa, disk, diskal ligament, and joint capsules, as 
well as the surface skin and mucosa may all be treated effectively with iontophoretic 

The amount of drug actually delivered to a given site and the depth of penetration 
depend primarily on the dose of the electrical current; the concentration of the drug and 
the total amount of the applied solution are less important (Oshima, Kashiki, Toyooka, 
Masuda & Amaha, 1994). 

The level of current applied to a site depends on the surface area (more current 
may be applied to a larger area) and to the individual's skin tolerance (Irsfeld et al., 1993; 
Lark & Gangarosa, 1 990). Increasing the electric current may decrease application time. 
However, this increases the risk of cutaneous burns. Lower levels are generally more 
comfortable (Oshima et al., 1994; Lark & Gangarosa, 1990) but take longer to apply. Up 
to 5 mA may be used if the individual can tolerate it (Byl et al., 1996). 

There are many advantages to using iontophoretic application instead of injection 
of medication. For example, the risks and inconveniences of injection or intravenous 
treatment, such as pain, tissue damage, and infection, are avoided. At the body surface, a 
high concentration of a medication can be administered into a lesion at a faster rate than 
with passive transport and without a detectable level of the drug appearing in the 
bloodstream or in other vital organs. On the other hand, if desired it allows delivery 
directly into the bloodstream and prevents the variation in absorption and metabolism seen 
in oral administration, and avoids the first pass elimination of a drug by the liver. A 

controlled, continuous delivery of a drug is ensured, preventing overdose and allowing 
rapid termination of delivery if necessary. Another advantage is that the body does not 
respond by immunologically sensitizing when medications such as antibiotics are delivered 
with iontophoresis. Additionally, iontophoresis may increase patient compliance and 
satisfaction and decrease anxiety through the avoidance of needles (Costello & Jeske, 
1995; Gangarosa, 1982; Kassanetal., 1996). 

A disadvantage of iontophoresis is the amount of time necessary for effectiveness. 
Another is the mild hyperemia, tingling, irritation, or burning that may occur on the skin 
surface. Because continuous, unidirectional current has an anesthetic effect (Byl et al., 
1996), these side-effects may occur without the sensation of pain and may heal slowly, 
especially if they occur under the cathode. However, it is possible to take measures to 
minimize the risk of such adverse effects. Patients with excessive susceptibility to the 
application of electrical currents, patients with cardiac pacemakers or other implanted 
devices, and patients with known sensitivity to the drugs being administered are not 
candidates for iontophoresis (Lark & Gangarosa, 1990). 

Effective treatment of many conditions in several fields has been described in the 
iontophoresis literature, including dermatology (e.g., application of antiviral agents to 
surface tissues, treatment of burns with antibiotics) (Kassan et al., 1996; Singh & Singh, 
1993), otolaryngology (e.g., for local anesthesia for simple surgeries of the tympanic 
membrane) (Costello & Jeske, 1995), and ophthalmology (treatment of eye lesions). 
Physical therapists have used iontophoresis of anti-inflammatory medications to treat a 
variety of musculoskeletal conditions (Costello & Jeske, 1995). Gangarosa (1988) has 
noted that iontophoresis is particularly well suited for treatment in dentistry, because many 

conditions are near the surface of the body. Applications include the treatment of 
hypersensitive dentin with fluoride, treatment of oral ulcers and herpes orolabialis, and 
delivery of local anesthetics for a variety of procedures, such as extraction of teeth, 
biopsies, anesthesia prior to injection or intravenous needle insertion, or as a substitute for 
injection for local anesthesia (Carlo, Ciancio & Seyrek, 1982; Costello & Jeske, 1995; 
Gangarosa, 1974; Gangarosa, 1981b; Gangarosa, 1982;Gangarosa, 1988). Lark and 
Gangarosa (1990) have reported the utility of iontophoretic delivery of local anesthetics or 
corticosteroids in treating temporomandibular joint and myofascial pain dysfunction 
syndromes. Acute muscle disorders, such as splinting and inflammation, disk disorders 
and inflammatory disorders of the joint, such as arthritis and capsulitis, are all responsive 
to medication delivery with iontophoresis. In addition, chronic mandibular hypomobilities, 
such as myostatic and myofibrotic contracture of the elevator muscles or capsular 
tightness, and growth disorders of the joint may benefit from iontophoretic administration 
of drugs. 

It has also been well established that iontophoresis is an effective method of 
achieving local anesthesia (Maloney et al, 1992). Moreover, iontophoresis provides 
significant advantages over traditional topical methods of anesthesia, which operate by 
passive diffusion over the concentration gradient across the skin (Singh & Singh, 1993). 
Studies have shown that topical cream anesthetics, such as a lidocaine/prilocaine mixture, 
can be reasonably effective; however, they require a considerably longer application time, 
and do not provide as complete an anesthetic effect as does iontophoresis (Greenbaum & 
Bernstein, 1994, Irsfeld, Klement & Lipfert, 1993). The depth of anesthesia achieved 
with iontophoresis has been shown to be comparable to that of injection, but the duration 

time is shorter (Russo, Lipman & Comstock, 1980). The shorter duration time is likely 
due to capillary dilation induced by iontophoresis, which increases regional blood flow and 
corresponding removal of the drug from the site (Kassan et al., 1996). Duration may be 
increased by using a solution containing epinephrine, a vasoconstrictor (Costello & Jeske, 
1995; Gangarosa 1981a). Because systemic administration of epinephrine causes 
sympathomimetic effects, it is recommended that iontophoretic application of epinephrine 
be avoided in persons with known sensitivity to epinephrine (Costello & Jeske, 1995). 

In sum, it is clear that researchers (e.g., Lark & Gangarosa, 1990) have attested to 
the clinical efficacy of iontophoresis as a method of treatment in certain TMD and MPDS 
conditions. In addition, the established use of iontophoresis in administering topical 
anesthetics, and the clear advantages of iontophoretically applied anesthetics over both 
simple topical application and injection, suggest that iontophoresis was an optimal choice 
for administration of local anesthesia for our research protocol. 

Design and Statement of Hypotheses 

This research was designed to examine the psychometric properties of the 
magnitude matching procedure using healthy subjects. Test-retest reliability was examined 
by administering the procedure to subjects on two separate occasions, with a range of 3 to 
8 days between sessions. The validity and clinical utility of the magnitude matching task 
were examined in a double-blind, placebo and no-treatment controlled paradigm. Subjects 
were randomly assigned into one of 3 conditions upon entry in the study, with an attempt 
made to obtain an equal number of male and female subjects in each condition. The 
experimental group (N=21) received iontophoretically applied lidocaine with epinephrine, 
whereas the control group (N=24) received placebo iontophoresis without lidocaine. The 

no-treatment group (N=20) was obtained using the data from session 1 of the test-retest 

For study #1, test-retest reliability, it was expected that the magnitude matching 
procedure would be highly reliable, as demonstrated by the reliability coefficient 
(Pearson's/ - ). 

For the validity study, it was anticipated that the experimental group, which 
received the anesthesia, would demonstrate decreased discriminability, shown by flatter 
(less steep) slopes than those of the other two groups. It was not expected that a 
significant difference would be found between slopes for the placebo and no-treatment 
control groups. It was hypothesized that both the experimental group and the placebo 
group would demonstrate lower Y-intercept values than the no-treatment control group, 
indicating a decrease in response bias, or a tendency to report the stimuli as less painful. 

A significant difference was predicted between the VAS scores for pain intensity 
and unpleasantness between the experimental and the two control groups, with the latter 
groups displaying higher scores on both scales. 



Subjects were recruited from the University of Florida undergraduate subject pool, 
the University of Florida Colleges of Dentistry and Health Professions. In addition, some 
subjects were University of Florida or Shands Teaching Hospital employees. All subjects 
were recruited in accordance with guidelines designated by the University of Florida 
Institutional Review Board and the Shands Teaching Hospital IRB. Some undergraduate 
subjects received course credit for their participation. 

Subjects were read a list of screening questions to ascertain that they met the 
criteria designed for subject safety (see Appendix A). Screening ensured that subjects had 
no skin conditions that would contraindicate the use of iontophoresis (e.g., discoloration 
or blemishes on the skin), and no past or current symptoms of pain in the jaws or face. 
Questionable subjects were given a palpation exam by one of the co-investigating dentists; 
subjects demonstrating pain or gross dysfunction were excluded from the study. Subjects 
also were screened for known sensitivity to epinephrine or a cardiac condition, factors 
which would contraindicate the use of iontophoresis and/or lidocaine with epinephrine. 


In study # 1, designed to establish test-retest reliability, subjects signed the 
informed consent form and completed a demographics questionnaire (see Appendix B) 
before undergoing the magnitude matching protocol described below. Between 3-8 days 


later, subjects returned to the laboratory for repetition of the magnitude matching. 
Following each session, subjects completed 10 cm, horizontal visual analog scales to 
indicate the intensity of overall pain and unpleasantness experienced during the procedure. 
At the end of the second testing session, subjects were given a full explanation of the 
purpose and hypotheses of the study. 

Subjects participating in Study #2, the validity study, were informed that they had 
been randomly assigned into one of two conditions, the anesthesia group or the placebo 
group. Subjects were told that the anesthesia device operates by a very mild and safe 
electrical current and that some subjects have reported feeling a "barely perceptible" 
sensation (Bushnell et al., 1991). Subjects completed the informed consent and 
demographics questionnaire. The possible adverse effects were described and subjects 
were given the opportunity to ask any questions about iontophoresis or the drugs being 

The experimental group was administered a dosage of 15 mA minutes of lidocaine 
2%, 1/100,000 epinephrine to the skin over the masseter muscle. This took an average of 
12-15 minutes per patient using a 1 .2 mA current. The treatment was performed using 
the Empi Du Pel iontophoresis unit (#LR64406), using Empi small size dispersive and 
active electrodes. This unit has been designed to automatically control the current and 
rate of dosage in accordance with the individual's skin resistance. The placebo control 
group underwent application of electrodes, without lidocaine and epinephrine, in the same 
location as the experimental group, and were given the same description of the sensation 
they may feel. However, the electrical current was not applied. The electrodes were in 
place for approximately 12 minutes, as with the experimental group. 

In both conditions, a separate trained examiner, blind to experimental condition, 
administered the magnitude matching protocol and the visual analog scales. Following 
completion of the study, subjects were given a full explanation of the purpose and 
hypotheses of the study. The subjects were instructed to use a moisturizer or hand lotion, 
preferably containing aloe, on the treated skin under both electrodes. 

Magnitude Matching Protocol 

In the magnitude matching paradigm, used to compare standardized aspects of 
pain ratings over time and between the experimental and control groups, 3 different levels 
(6, 7, and 8 foot pounds per square inch) of manual pressure were applied using a hand- 
held pressure algometer (Pain Diagnostics, Inc). Lines of 3 different lengths (70, 85 and 
100 mm), drawn on 5x7 inch index cards, were presented as standardization stimuli. 
Subjects were asked to rate pressure stimuli and line lengths on the same generic number 
scale, where indicates no discomfort or pain and no length, and where the top end of the 
scale is open. Subjects are instructed to use larger numbers for longer lines and greater 
pain or discomfort, and smaller numbers for shorter lines and less pain or discomfort. The 
instructions for this procedure may be found in Appendix C. 

Manual pressure was applied at the insertion of the masseter muscle, one 
centimeter superior and anterior to the angle of the mandible, as described in the 
instructions for palpation in the Research and Diagnostic Criteria guidelines for assessment 
of temporomandibular disorders (Dworkin & LeResche, 1992). Training in location of 
this point was provided to the experimenter by the investigating dentist, a specialist in the 
assessment and treatment of facial pain. This point was marked in ink in order to ensure 

consistency in site of application. Pressure was increased over one second, then held for 
two seconds. The importance of standardized rate of application has been emphasized by 
several authors (e.g., Jensen et al., 1986). Using a pad size approximately equal to one 
square centimeter circular area, ratings were assessed at 6, 7 and 8 foot pounds per square 
inch; pilot data collected using an up-down transformation sequence procedure ( Wetherill 
& Levitt, 1965) on 12 nonpatient subjects (6 female, 6 male) revealed an average pain 
threshold of 7 foot pounds per square inch. 

Subjects were given two practice trials, applied to the right wrist bone, using 
pressure stimuli of 1 and 3 foot pounds per square inch. Once it was clear that the subject 
understood the procedure, the 3 pressure levels were delivered in a randomized sequence 
of groups of 3, with no level repeated until all other levels have been presented. 
Presentation of the 3 pain stimuli was alternated with the similarly randomized display of 3 
line lengths, for a total of 9 pressure ratings and 9 ratings of line length. 

Because degree of muscle tension in the jaw has been shown to influence the 
pressure-pain threshold (McMillan & Lawson, 1994), subjects were asked to place their 
teeth together without clenching their jaw. For ease of administration, data was collected 
from the right side of the face. Reid et al. (1994) reported that in control subjects, 
pressure-pain thresholds differed among various sites, but no difference between the two 
sides of the face was found. Furthermore, some authors have reported a significantly 
higher threshold on the right side, seen only in right handed subjects; such effects were 
not found in left handed or ambidextrous subjects (Jensen, Rasmussen, Pedersen, Lous & 
Olesen, 1992). 

The magnitude matching paradigm was used to obtain 2 indices, the slope and 
intercept of standardized regression lines. The slope represents ability to discriminate 
between pressure stimuli, and the intercept represents response bias. Derivation of these 
indices is described in the Results section below. 
Visual Analog Scales 

Following the magnitude matching procedure, subjects were asked to rate overall 
painfulness and overall unpleasantness on 2 horizontal lines of 1 cm length drawn on a 
half-sheet of 8x10 inch paper. Endpoints were designated as "not at all 
painful/unpleasant" and "the worst pain/most unpleasant imaginable." Subjects were 
instructed to indicate degree of pain/unpleasantness by drawing a vertical line at the 
appropriate point on the scale. Price & Harkins (1988) describe the many advantages of 
visual analog scales in the assessment of pain intensity, including evidence for power 
functions and ratio scale data, ease of administration, relatively bias-free results, validity 
and reliability, and sensitivity to pain-reducing interventions. 


All analyses were conducted using the SPSS 7.5 for Windows 95 statistical 
program (Norusis/SPSS Inc, 1995). The data were first examined for violation of 
assumptions. Through examination of histograms plotted for each group, outliers were 
identified by distance from the mean. An attempt was made to ascertain the reason for 
anomalous results given by 7 subjects. Equipment failures, scaling in the wrong direction, 
and failing to scale consistently or on one common scale were the cause of outliers. Due 
to a differential outlier rate in the experimental group (4 subjects), which dropped the 
number of cases to below the goal of at least 20 subjects per condition, the decision was 
made to drop all outliers and to replace the subjects in the experimental group. 
Replacement ensured adequate power in detecting significant differences. 

Frequencies and percentages and, for noncategorical variables, ranges and means, 
for the demographics of sex, age, ethnicity, income, education, marital status, and 
occupation are shown in Tables 1 and 2, respectively. A one-way ANOVA confirmed that 
there were no significant differences between the experimental, placebo control and no- 
treatment control groups on any demographic variables. 

For both the test-retest reliability and validity studies, calculation of the indices 
derived from the magnitude matching paradigm were based on the methods described by 
Stevens and Marks (1980), Duncan et al. (1988), Feine and colleagues (1991), and Fuller 
and Robinson (1995). The mean rating of each of the 3 pressure levels was first calculated 



Table 1 

Frequencies and Percentages for Demographic Variables 















African American 









East Indian 

























> 100,000 



Student Status 










Marital Status 

















Table 2 

Ranges, Means and Standard Deviations for Noncategorical Demographic and Visual 

Analog Scale Variables 











VAS Pain 





VAS Unpleasantness 





Retest Pain 





Retest Unpleasantness 





for each subject. Then, in order to standardize these mean ratings for each subject, each 
subject's mean rating was divided by that subject's grand mean rating of all of the control 
stimuli (lines). Each subject thus generated 3 standardized values, one for each level of 
algometer pressure. Regression equations were calculated for each condition, with the 
standardized ratings as the dependent variable and the pressure level as the independent 
variable. These regressions yielded a slope and intercept value for the magnitude 
matching task. The slope and intercept values were then used to examine test-retest 
reliability and to test for differences between the experimental (anesthesia), placebo 
control (placebo anesthesia), and no-treatment control (Time 1 of the test-retest reliability 
study) groups. 

To examine test-retest reliability, Pearson's r was calculated for the slopes and 
intercepts between Session 1 and Session 2. The results of these analyses provided 
support for moderate test-retest reliability of the slope, but not the intercept, over a 3-8 
day period. The correlation between the slopes across the two sessions was .71 (p < .01). 
For the intercepts, a nonsignificant r of .44 was obtained. 

To test for slope differences between the 3 groups in the validity study, a one-way 
ANOVA was conducted. In addition, pairwise comparisons had been planned a priori. 
Intercept differences were also tested with a one-way ANOVA. The means and standard 
deviations for each of the 3 groups and the results of the ANOVA tests are presented in 
Tables 3 and 4. The results indicate that there were no significant differences in slope 
(discriminability) between the 3 groups, F (2, 62)=. 13, p_ = .88. Pairwise comparisons 
were therefore not conducted. However, significant intercept (response bias) differences 
were obtained: F (2, 62) = 8.03, p = .001 . The planned pairwise comparisons with Tukey 


Table 3 

Means and Standard Deviations for One- Way ANOVA - Intercept 

Group Mean S.D. 

Placebo Control* 1.06 .27 

Treatment .74 .24 

No-treatment Control* .93 .28 
F(2, 62) = 8.03, p. = .001 

Table 4 

Means and Standard Deviations for One-Way ANOVA - Slope 

Group Mean S.D. 

Placebo Control .14 .09 

Treatment .15 .07 

No-treatment Control .15 .08 
F(2, 62) =13, p =.88 

correction revealed that both the placebo control and no-treatment control groups were 
significantly different from the experimental group. However, the two control groups 
were not significantly different from each other. 

The VAS mean ratings and standard deviations for overall pain and unpleasantness 
are shown for each group in Table 2. A one-way ANOVA found no significant 
differences on these ratings between any of the 3 conditions. For overall pain, F (2, 59) = 
1 .34, p = .27. For unpleasantness, F (2, 59) = .66, p - .52. 


The aim of this research was to examine the psychometric properties, validity and 
test-retest reliability, of a magnitude matching task. The protocol employed the 
methodology described by Stevens and Marks (1980). A pressure algometer (Ohrbach & 
Gale, 1 989a) was used to provide a clinically relevant pain stimulus, emphasized by Fuller 
and Robinson (1995). Magnitude matching techniques yield regression lines which 
provide separate values for two components of pain threshold: discriminability (slope) 
and response bias (intercept). It was hypothesized that the slope and intercept would 
show high reliability between the two testing sessions. Moderate test-retest reliability over 
a 3-8 day time period was shown for the discriminability component. However, poor test- 
retest reliability was observed for response bias for the same time period. It was also 
expected that the magnitude matching procedure would be sensitive to a known analgesic, 
iontophoretically administered lidocaine with epinephrine. Specifically, the treatment 
group was expected to demonstrate a reduction in discriminability, shown by flatter 
slopes, as compared with the two control groups. In addition, both the experimental and 
placebo group were hypothesized to show a decrease in response bias, indicated by lower 
intercept values, as compared with the no-treatment control group. The results of the 
validity test indicate that the magnitude matching procedure was indeed sensitive to the 
group differences induced by the iontophoretically applied lidocaine and epinephrine 
anesthetic. This is evidenced by the significant differences found between the 3 groups on 


the intercept values. Both the placebo and no-treatment control groups were significantly 
different from the anesthesia group, but not from each other, with both control groups 
tending to report more pain. 

Test-Retest Study 

The data indicate that the study participants were able to discriminate between the 
3 pressure levels (6, 7 and 8 foot pounds per square inch) equally well on both testing 
occasions. The lack of reliability of the intercept component suggests that unknown 
factors affected the subjects' response bias in an inconsistent way across time. Possible 
mediators may include learning or state factors such as mood. 

In the present study, subjects were tested while sitting in a dental chair in a clinical 
examination room. As part of the consent process, they were informed that some of the 
pressures they would be receiving may produce temporary discomfort or pain. These 
aspects of the experimental situation, along with unfamiliarity with the task, may have 
contributed to a state of increased anxiety that was not operative during the second testing 
session. Other research has suggested that task instructions may be anxiety-inducing and 
that the resulting anxiety can affect task performance and signal detection theory 
parameters (Dougher, 1979; Malow, 1981; Schumacher & Velden, 1984; Weisenberg, 
Anram, Wolf & Raphali, 1984). 

In their research on pressure-pain threshold with the pressure algometer, Jensen et 
al. (1986) found a slight but significant increase in PPT over the course of 5 weekly 
testing sessions. The authors suggest a reduction in anxiety with increased familiarity with 
the procedure or a learning effect as possible mediating factors. Although the PPT does 
not yield separate values for discriminability and response bias, the authors' observed 

increase in PPT over time, if due to anxiety, would be consistent with an increase in 
response bias, or a lesser tendency to label the test stimuli as painful. 

Validity Study 

Unexpectedly, it was found that the placebo group tended to report the most pain, 
as demonstrated by the intercept values (see Table 3). This higher value suggests that 
factors associated with the application of the placebo condition contributed to increased 
response bias in that group. In contrast, such factors did not have consistent effects on the 
other two groups. For example, the identical instructions given to the placebo and 
treatment groups included the potential risks and discomforts of the electrode pads, 
lidocaine and epinephrine, and iontophoresis. It is possible that factors such as anxiety 
influenced the response bias in these groups. The link between anxiety and pain has been 
well established by early researchers (Melzack & Dennis, 1971; Rollman, 1977; 
Sternbach, 1978). Elevated anxiety levels have been associated with reduced pain 
tolerance and lowered pain thresholds in medical, psychiatric and nonpatient samples 
(Malow, West & Sutker, 1987). It is generally assumed that greater anxiety contributes to 
a greater pain reaction to painful stimuli (Malow, 1981). State anxiety has also been 
correlated with affective and evaluative, but not sensory, dimensions on the McGill Pain 
Questionnaire. These results would be consistent with differences in response bias. 

In the present study, the anesthetic may have counteracted the effects of anxiety in 
the treatment group whereas the placebo control group may have continued to experience 
these effects. As a result, their response bias values were elevated in comparison to the 
treatment group. Furthermore, the participants in the no-treatment control group were 
aware that the risks described above were not applicable to them; thus their response bias 

may have remained unaffected by anxiety or other factors which influenced the other 
groups. If anxiety did indeed mediate these results, the observed effect for response bias 
may actually be strengthened. Response bias represents subjects' willingness to report 
pain. Thus, subjects in the placebo group did not show a tendency to inhibit their report 
of pain; anxiety may have in fact heightened their pain. It is important to note, however, 
that in the present study, anxiety during the procedure was neither assessed nor 

In contrast with the intercept differences found, and with the hypothesis, no 
significant differences were found between the 3 groups in discriminability, as illustrated 
by the parallel slopes in Figure 2. Thus, it appears that the experimental manipulation did 
not affect this variable; all groups were apparently able to discriminate between the 3 
levels of pressure equally well. Figure 2 also illustrates the standardized pressure variables 
at each pressure level. 

Although the present study was not designed to investigate the mechanisms of the 
anesthetic used as the treatment, the known mechanisms of local anesthetics may account 
for the observed group differences in response bias versus discriminability. It is possible 
that the experimental manipulation was effective enough to differentially influence the 
groups' response bias, resulting in differences in the report of pain, without substantially 
altering their ability to discern the differences in pressure level. In other words, subjects in 
each group could still discriminate the pressure levels, but they differed in whether or not 
they labeled the stimuli as painful. 

Several factors are known to contribute to differential blocking of the various 
types of nerve fibers. For example, the minimum blocking of an anesthetic, defined as the 


No Treatment! 

6 7 8 

Pressure Level 

Figure 1 . Standardized Mean Pain Ratings for 6, 7, and 8 Pounds 

"drug concentration that just halts impulse traffic" (de Jong, 1996, 150) varies with nerve 
fiber diameter and concentration of voltage-gated sodium ion channels. These channels 
act as binding sites for lidocaine and other local anesthetics. In addition, nerve blockade is 
frequency dependent, meaning that opportunities for the local anesthetic to bind to the 
sodium channels are enhanced when the frequency of stimulation is increased. The faster 
the nerve is made to fire, the more complete the resulting nerve block. As a result, fibers 
such as nociceptive and sympathetic nerves, which carry rapidly flowing impulses, require 
a less concentrated local anesthetic to disrupt impulse propagation than do large motor 
fibers. Thus, the latter may remain functional while A-delta and C fibers, carrying pain- 
related impulses, are inhibited (de Jong, 1996). In general, local anesthetics first result in a 
loss of sympathetic function, then loss of pinprick sensation, followed by touch and 
temperature discrimination, with loss of motor function affected last (Barash, Cullen, & 
Stoelting, 1997; de Jong, 1996). 

Lidocaine, in clinical doses, completely inhibits depolarization. The addition of 
epinephrine, a vasoconstrictor, decreases the rate of vascular absorption, allowing more 
anesthetic molecules to reach the nerve membrane and improving both the depth and 
duration of anesthesia (Berde & Strichartz, 1994). 

However, because of the differential nerve block factor described above, an 
individual treated with lidocaine (and epinephrine) may still perceive touch and pressure 
after pain sensations have been eliminated (de Jong, 1996). Therefore, it is highly possible 
that discriminability between the 3 levels of pressure used in the present study remained 
unaffected on the basis of the mechanisms of action of lidocaine. Although Lineberry and 
Kulics (1978) reported that in rhesus monkeys, subcutaneous injection of lidocaine leading 

to a partial local block of peripheral nerves resulted in a decrease of discriminability 
between noxious intensities of electrical stimuli (which presumably are affected by 
lidocaine quicker than are pressure sensations) in a signal detection study, it appears that 
the differential effects of lidocaine on discriminability in humans are undocumented in the 

A considerable body of earlier signal detection research illustrates the differential 
effects of drugs on response bias versus discriminability. For example, Chapman, Murphy 
and Butler (1973) found that a 33% nitrous oxide solution affected a change in both 
discriminability of certain heat stimuli and response bias. In contrast, diazepam was found 
not to affect sensitivity or response bias relative to placebo (Chapman & Feather, 1973). 
In studies of placebo alone, it was found that a placebo analgesic increased response bias, 
but not discriminability, to pain, heat and warmth (Clark, 1969). Feather, Chapman and 
Fisher (1972) obtained similar results. The signal detection research employed 
systemically, rather than cutaneously delivered drugs, which would likely hold greater 
influence over cognitive tasks such as discrimination. However, it is possible that 
lidocaine also affects discriminability and response bias differentially. Both the observed 
differences in response bias and the lack of differences in discriminability may be due to 
the effects of the lidocaine itself or to other factors. 

Other reasons why the results of the present study did not yield group differences 
in discriminability may involve methodological issues. For example, Green and Swets 
( 1 966) point out that in classic psychophysics studies, performance may stabilize only after 
many sessions. When only a few trials are conducted, data is then collected during the 
period when subject performance is at its maximal level of fluctuation, and both 

discrimination and response bias may change during a single testing session. It is possible 
that not enough trials were administered to reveal group differences. Fluctuation may also 
have affected the results of the test-retest study if performance was not yet stabilized. 

However, repeated trials in the present study would have compromised the 
ecological validity to the clinical situation. In the assessment of facial pain in patient 
samples, it would not be possible to administer many pressure trials without aggravating 
existing pain conditions. Because the magnitude matching protocol is intended for 
eventual use with pain populations, clinical relevance was prioritized. 

Visual Analog Scales 

It was hypothesized that significant differences would be found on overall pain and 
overall unpleasantness as measured by the visual analog scales. However, no differences 
were found among the 3 groups on either of these variables. It is possible that the VAS 
measures were not sensitive to group differences, perhaps due to their greater face validity 
and subjects' desire to minimize their experience of pain and unpleasantness in front of the 


Unique features of the present research included the use of a magnitude matching 
task using a clinically relevant pain stimulus, a pressure algometer. Both magnitude 
matching and pressure algometry are believed to represent improvements over more 
traditional means of pain assessment, such as other scaling or threshold measures using 
experimental pain stimuli, that have unclear relevance to clinical pain. In addition, 
iontophoresis was employed to facilitate the action of topically applied lidocaine with 
epinephrine, a standard treatment in many clinical settings. Test-retest reliability was 
examined over a 3-8 day period and validity was investigated with a double-blind, placebo 
and no-treatment control experimental design. 

Test-retest reliability was found to be moderate for the discriminability index but 
poor for response bias over time. Subjects were apparently able to scale consistently 
across the testing sessions but their performance was affected by nonsensory factors. 
Anxiety related to the task and laboratory conditions, which diminished with familiarity by 
the second session, is a possible explanatory factor. 

Discriminant validity for the magnitude matching task using pressure algometry 
was demonstrated by the sensitivity of the procedure to a known analgesic, 
iontophoretically applied lidocaine with epinephrine. Group differences were seen 
between each of the control groups and the treatment group on the measure of response 
bias. These results are consistent with the hypotheses. However, in contrast to 


hypotheses, no group differences were observed in discriminability. These data suggest 
that, relative to both placebo anesthesia and no treatment at all, subjects in the anesthesia 
group tended to report lower levels of pain. All groups were able to discriminate between 
the 3 levels of pressure. These results may be explained by the differential mechanisms of 
action of lidocaine on discriminability or by a combination of other factors such as anxiety 
or methodological constraints. 

Visual analog scales for overall pain and aversiveness related to the magnitude 
matching protocol was assessed for comparative purposes. Although group differences 
were expected on both of these scales, none were found. It is possible that these measures 
were more face valid and less sensitive to differences than was the magnitude matching 

An aim of the present research was to examine the psychometric properties of the 
magnitude matching using healthy subjects, as a precursor to research with patient 
populations. Specifically, the use of the algometer to deliver pressures to the masseter 
muscle was designed with considerations for the assessment of facial pain. The fact that 
support was found for the psychometrics of the protocol suggests that magnitude 
matching methodology may be applied to this and other pain populations. Because it is 
sensitive to an analgesic, magnitude matching may be able to in distinguish between 
patient and nonpatient samples. In addition, the results suggest that magnitude matching 
may be useful as an outcome measure for a variety of treatments for clinical pain. 
Clinically, the methodology yields more information than do more traditional assessment 
measures, such as other scaling techniques or pain thresholds. Of interest is the fact that a 
plot of the log transformed standardized pressure variable for each condition revealed a 

linear relationship (see Figure 2). These results demonstrated that subjects were able to 
discriminate between the levels of pressure while accurately scaling in a ratio level manner. 
Thus, the magnitude matching procedure shows clear psychometric advantage over more 
traditional methods of scaling. In addition, the separation of response bias from 
discriminability may in the future contribute to matching of treatment to patient 
characteristics or the prediction of treatment success. 

Further research is necessary to delineate factors that affected the response bias 
dimension in the present studies. In both the reliability and validity studies, it appeared 
that anxiety due to the experimental situation played a role in the study results. However, 
because anxiety was not assessed or measured, its role is speculative. The effects of 
anxiety and of other factors, such as learning, and the extent to which these can be 
experimentally manipulated, should be examined. Similarly, future research needs to 
identify factors that may have influenced the lack of demonstrated group differences in 
discriminability. For example, the choice of lesser, or further spaced pressures, may 
positively influence subject performance. More specific instructions, training, and the 
addition of practice sessions may also increase discriminability. Related to these factors is 
the issue of stability of performance, which should also be examined using more traditional 
psychophysical methods. 

In general, the results of the present research are encouraging. The methodology 
may applied to a variety of experimental and clinical situations and holds promise for the 
clinical assessment of pain conditions and treatment outcome, as well as for future 




A treatment 
















</> 0.04 






log at each pressure level 


Figure 2. Log Transformed Variables 


Subject's Name : Date : 

ID.# : 

Skin Complexion 

1 . Do you have very sensitive skin? Yes (Explain) No 

2. Do you sunburn easily? Yes No 

3 . Is your skin very fair, with many freckles? Yes No 

4. What is your natural hair color? 

5. Do you have a skin reaction to band-aids or other adhesive tapes? Yes No 

6. Do you have significant facial scarring due to acne or blemishes? Yes No 

7. Do you have any birthmarks or other skin discoloration on your face? Yes No 

8. If male, do you have a beard? Yes No 

Facial Pain 

1. Do you have recurrent or chronic pain in your jaws or face? Yes No 
(If yes, explain ) 

2. Have you ever had recurrent or chronic pain in your jaws or face? Yes No 
(If yes, explain ) 


1 . Do you have any heart problems? Yes No 
(If yes, explain ) 

2. Do you have a cardiac pacemaker? Yes No 


1 . Do you have a known drug allergy to lidocaine? Yes No 

2. Do you have a known drug allergy to epinephrine? Yes No 

3. If female, is there any chance that you could be pregnant? Yes No 

Is subject suitable for participation in study? Yes No 

Randomized to which study? Test-Retest Validity (Experimental or Control?) 



Please answer aU questions. 

Name: ID.# Date: 

Age: Sex: O-Male 1 -Female 

Ethnicity (Circle one): 

1 -Caucasian 2-African- American 3-Hispanic 4- Asian 5-East Indian 
6-Bi-racial (list) 7-Other (list) 

Education (Circle highest completed): 

High School/GED College: 12 3 4 Graduate School: 12 3 4 5+ 

Graduate Program 

Are you currently a student? 0-No 1 -Yes, Undergraduate 2- Yes, Graduate 


Marital Status (Circle one): 

1 -Single 2-Married 3 -Separated/Divorced 4-Widowed 

Total family income (circle one): 

1 -Under $15,000 2-$ 15,000-30,000 3-$3 1,000-45,000 

4-$46,000-60,000 5-$6 1,000-75,000 6-$76,000- 100,000 

7-Over $100,000 

Do you have, or have you ever had, a problem with pain in your jaw? 0-No 1-Yes 
In another location? 0-No 1-Yes 

Does anyone in your family have a problem with chronic pain? 0-No 1-Yes 

If yes, who? 



Instructions to subjects: 

"I'm going to ask you to rate the sensation of pressure applied to your jaw with this 
pressure device, and also to rate the length of lines presented on these cards. We will 
repeat these ratings nine times, three times for each line and pressure. Each rating will 
take approximately three seconds. This test is to get an accurate understanding of the pain 
and sensitivity of your face. 

Judge the sensation and length of the lines on the same number scale. Zero on the 
scale represents no sensation, and no length. You can use numbers as large as you want, 
but I want you to use larger numbers for longer lines and more painful or uncomfortable 
sensations, and smaller numbers for shorter lines and less painful or uncomfortable 

Let's practice." 

1 Present pressure of 1 foot pound per square inch on wrist bone. Ask for rating. 

2 Present card # 3. Ask for rating. 

Present pressure of 3 foot pounds per square inch on wrist bone. Ask for rating. 
3. Present card #1 . Ask for rating. 



Al-Hassan, H.K., Ismail, A I , & Ash, M.M. (1986). Concerns of patients seeking 
treatment for TMJ dysfunction. Journal of Prosthetic Dentistry, 56 , 217-221 . 

Barash, P.G., Cullen, B.F., & Stoelting, R.K. (1997). Handbook of clinical anesthesia, 
third edition . Philadelphia: Lippincolt-Raven. 

Berde, C.B., & Strichartz, G.R. (1994). Local anesthetics. In R.D. Miller (Ed), 
Anesthesia, fourth edition (pp. 489-509). New York: Churchill Livingstone. 

Brown, F.F., Robinson, ME, Riley, J.L., & Gremillion, HA. (1996). Pain severity, 
negative affect, and microstressors as predictors of life interference in TMD patients. 
Journal of Craniomandibular Practice, 14 (1), 63-70. 

Bushnell, M.C., Marchand, S., Tremblay, N., & Duncan, G.H. (1991). Electrical 

stimulation of peripheral and central pathways for the relief of musculoskeletal pain. 
Canadian Journal of Physiology and Pharmacology, 69 , 697-703. 

Byl, N., Zellerbach, L.R., & Pfalzer, LA. (1996). Systemic issues and skin conditions: 
Wound healing, oxygen percutaneous drug delivery, burns, and desensitized skin. In 
R.S. Myers (Ed.), Saunders manual of physical therapy practice (pp. 585-623). 
Philadelphia: W.B. Saunders Company. 

Carlo, G.T., Ciancio, S.G., & Seyrek, S.K. (1982). An evaluation of iontophoretic 
application of fluoride for tooth desensitization. Journal of the American Dental 
Association. 105 . 452-454. 

Costello, C.T., & Jeske, AH. (1995). Iontophoresis: Applications in transdermal 
medication delivery. Physical Therapy. 75 . 554-563. 

deJong, R.H. (1996). Local anesthetics in clinical practice. In S.D. Waldman & A.P. 
Winnie (Eds), Interventional pain management ("pp. 151-166). Philadelphia: W.B. 
Saunders Company. 

de Wijer, A., Lobbezoo-Scholte, A.M., Steenks, M.H., & Bosman, F. (1995). Reliability 
of clinical findings in temporomandibular disorders. Journal of Orofacial Pain. 9 . 181- 



Dougher, M.J. (1979). Sensory decision theory analysis of the effects of anxiety and 
experimental instructions on pain. Journal of Abnormal Psychology, 88 (2), 137-144. 

Duncan, G.H., Feine, J.S., Bushnell, M.C., & Boyer, M. (1988). Use of magnitude 
matching for measuring group differences in pain perception. In R. Dubner, G.F. 
Gebhart, & M.R. Bond (Eds), Proceedings of the Vth World Congress on Pain (pp, 
383-390). New York: Elsevier Science Publishers. 

Dworkin, S. F. (1995). Personal and societal impact of orofacial pain. In JR. Fricton & 
R. Dubner (Eds.), Orofacial pain and temporomandibular disorders (Vol. 21, pp. 15- 
32). New York: Raven. 

Dworkin, S.F., & LeResche, L. (1992). Research diagnostic criteria for 

temporomandibular disorders: Review, criteria, examinations and specifications, 
critique. Journal of Craniomandibular Disorders. Facial and Oral Pain. 6 , 301-355. 

Dworkin, S.F., LeResche, L., & DeRouen, T (1988). Reliability of clinical measurement 

in temporomandibular disorders Clinical Journal of Pain. 4 , 89-99. 

Feine, J., Bushnell, M., Miron, D., & Duncan, G. (1991). Sex differences in the 
perception of noxious heat stimuli. Pain. 44 . 255-262. 

Fuller, A. K., & Robinson, ME. (1995). Perceptual differences between patients with 
chronic low back pain and healthy volunteers using magnitude matching and clinically 
relevant stimuli. Behavior Therapy. 26 . 241-253. 

Gangarosa, LP. (1974). Iontophoresis for surface local anesthesia. Journal of the 
American Dental Association. 88 . 125-128. 

Gangarosa, LP. (1981a). Defining a practical solution for iontophoretic local anesthesia 
of skin. Methods and Findings in Experimental and Clinical Pharmacology. 3 (2).83- 

Gangarosa, LP. (1981b). Iontophoretic application of fluoride by tray techniques for 
desensitization of multiple teeth. Journal of the American Dental Association. 102 . 

Gangarosa, LP. (1982). Iontophoresis in dental practice . Chicago: Quintessence 
Publishing Company, Inc. 

Gangarosa, L.P. (1988). Electrical medication (iontophoresis): A modality for expanding 
dental practice with new therapies. General Dentistry . Sept. -Oct., 402-404. 

Gangarosa, LP., & Mahan, P.E. (1982). Pharmacologic management of TMJ-MPDS. 

Ear. Nose and Throat. 61 . 30-41. 


Gangarosa, LP., Ozawa, A., Onkido, M., Shimomura, Y , & Hill, T.M. (1995). Journal 
ofDermatology.22 (in. 865-875. 

Gracely, R.H., & Reid, K.I. (1995). Orofacial pain measurement. In JR. Fricton & R. 
Dubner (Eds), Orofacial pain and temporomandibular disorders (Vol. 21, pp.1 17- 
143). New York: Raven. 

Graff-Radford, SB., Reeves, J.L., Baker, R.L., & Chiu, D. (1989). Effects of 

transcutaneous electrical nerve stimulation on myofascial pain and trigger point 
sensitivity. Pain, 37 . 1-5. 

Green, C.S., & Marbach, J.J. (1982). Epidemiological studies of mandibular dysfunction: 
A critical review. Journal of Prosthetic Dentistry. 48 . 184-190. 

Green, DM, & Swets, J.A. (1966). Signal detection theory and psvchophysics . New 
York: Wiley. 

Greenbaum, S.S., & Bernstein, E.F. (1994). Comparison of iontophoresis of lidocaine 
with a eutectic mixture of lidocaine and prilocaine (EMLA) for topically administered 
local anesthesia. Journal of Dermatology and Surgical Oncology. 20 . 579-583. 

Heft, M.W., & Parker, S.R. (1984). An experimental basis for revising the graphic rating 
scale for pain. Pain. 19 . 152-161. 

Irsfeld, S., Klement, W., & Lipfert, P. (1993). Dermal anesthesia: Comparison of EMLA 
cream with iontophoretic local anesthesia. British Journal of Anaesthesia. 71 . 375- 

Jaeger, B., & Reeves, J.L. (1986). Quantification of changes in myofascial trigger point 
sensitivity with the pressure algometer following passive stretch. Pain. 27 . 203-210. 

Jensen, K., Anderson, HO., Olesen, J., & Lindblom, U. (1986). Pressure-pain threshold 
in human temporal region. Evaluation of a new pressure algometer. Pain. 25 . 313- 

Jensen, R., Rasmussen, B.K., Pedersen, B., Lous, I., & Olesen, J. (1992). Cephalic 

muscle tenderness and pressure pain threshold in a general population. Pain. 48 197- 

Kassan, D.G, Lynch, A.M., & Stiller, M.J. (1996). Physical enhancement of 
dermatologic drug delivery. Iontophoresis and phonophoresis. Journal of the 
American Academy of Dermatology. 34 . 657-666. 


Lark, MR., & Gangarosa, LP. (1990). Iontophoresis: An effective modality for the 
treatment of inflammatory disorders of the temporomandibular joint and myofascial 
pain. Journal of Craniomandibular Practice, 8 (2). 108-119. 

List, T , Helkimo, M, & Falk, G. (1989). Reliability and validity of a pressure threshold 
meter in recording tenderness in the masseter muscle and the anterior temporalis 
muscle. Journal of Craniomandibular Practice, 7 (3), 223-229. 

List, T., Helkimo, M, & Karlsson, R. (1991). Influence of pressure rates on the 
reliability of a pressure threshold meter. Journal of Craniomandibular Disorders, 
Facial and Oral Pain. 5 . 173-178. 

Maloney, J.M., Bezzant, J.L., Stephen, R.L., & Petelenz, T.J. (1992). Iontophoretic 
administration of lidocaine anesthesia in office practice: An appraisal. Journal of 
Dermatology and Surgical Oncology, 1 8 , 937-940. 

Malow, R.M. (1981). The effects of induced anxiety on pain perception: A signal 
detection analysis. Pain, 11 , 397-405. 

Marks, L.E (1991). Reliability of magnitude matching. Perception and Psvchophysics . 
49(1), 31-37. 

Marks, L.E., Szczesiul, R., & Ohlott, P. (1986). On the cross-modal perception of 

intensity. Journal of Experimental Psychology: Human Perception and Performance, 
12(4), 517-534. 

McMillan, AS., & Blasberg, B. (1994). Pain-pressure threshold in painful jaw muscles 
following trigger point injection. Journal of Orofacial Pain. 8 (4). 384-390. 

McMillan, AS., & Lawson, E.T. (1994). Effect of tooth clenching and jaw opening on 
pain-pressure threshold in the human jaw muscles. Journal of Orofacial Pain. 8 (3). 

Ohrbach, R., & Gale, E.N. (1989a). Pressure pain thresholds, clinical assessment, and 
differential diagnosis: Reliability and validity in patients with myogenic pain. Pain. 
39, 157-169. 

Ohrbach, R., & Gale, E.N. (1989b). Pressure pain thresholds in normal muscles: 

Reliability, measurement effects, and topographical differences. Pain. 37 . 257-263. 

Oshima, T., Kashiki, K., Toyooka, H, Masuda, A., & Amaha, K. (1994). Cutaneous 
iontophoretic application of condensed lidocaine. Canadian Journal of Anesthesia. 
4_i(8), 677-679. 


Price, D. (1988). Psychological and Neural Mechanisms of Pain . New York: Raven 

Price, D., & Harkins, S.W. (1992). Psychophysical approaches to pain measurement and 
assessment. In DC. Turk & R. Melzack (Eds), Handbook of pain assessment (pp. 
1 1 1-134). New York: Guilford Press. 

Reeves, J.L., Jaeger, B , & Graff-Radford, S.B. (1986). Reliability of the pressure 
algometer as a measure of myofascial trigger point sensitivity. Pain, 24 , 3 13-321 . 

Reid, K.I., Dionne, R.A., Sicard-Rosenbaum, L., Lord, D., & Dubner, R.A. (1994). 
Evaluation of iontophoretically applied dexamethasone for painful pathologic 
temporomandibular joints. Oral Surgery, Oral Medicine, and Oral Pathology. 77 . 

Reid, K.I., Gracely, R.H., & Dubner, R.A. (1994). The influence of time, facial side, and 
location on pain-pressure thresholds in chronic myogenous temporomandibular 
disorder. Journal of Orofacial Pain. 8 (3). 258-265. 

Rollman, G.B. (1977). Signal detection theory measurement of pain: A review and 
critique. Pain. 3 . 187-211. 

Russo, J., Lipman, AG., & Comstock, T.J. (1980). Lidocaine anesthesia: Comparison of 
iontophoresis, injection, and swabbing. American Journal of Hospital Pharmacology. 

Schumacher, R., & Velden, M. (1984). Anxiety, pain experience, and pain report: A 
signal detection study. Perceptual and Motor Skills. 58 . 339-349. 

Singh, S., & Singh, J. (1993). Transdermal drug delivery by passive diffusion and 
iontophoresis: A review. Medicinal Research Reviews. 13 (5). 569-621. 

Speculand, B , & Goss, A.N. (1985). Psychological factors in temporomandibular joint 
dysfunction pain. International Journal of Oral Surgery. 14 . 131-137. 

Stevens, J.C., & Marks, L.E. (1980). Cross-modality matching functions generated by 
magnitude estimation. Perception and Psvchophvsics. 27 (5). 379-389. 

Stockstill, J., Gross, A., & McCall, WD. (1989). Interrater reliability in masticatory 
muscle palpation. Craniomandibular Disorders. Facial and Oral Pain. (3), 143-146. 

Svensson, P., Arendt-Nielsen, L., Nielsen, H , & Larsen, J.K. (1995). Effect of chronic 
and experimental jaw muscle pain on pain-pressure thresholds and stimulus-response 
curves. Journal of Orofacial Pain. 9 . 347-356. 


Weisenberg, M, Anram, O., Wolf, Y., & Rapheli, N. (1984). Relevant and irrelevant 
anxiety in the reaction to pain. Pain, 20 , 371-383. 

Wetherill, G.B., & Levitt, H. (1965). Sequential estimation of points on a psychometric 
function. The British Journal of Mathematical and Statistical Psychology. 18 (Pt 1). 

Felicia F. Brown received an undergraduate degree in psychology and 
anthropology from Tufts University in 1990. She was employed as a neuropsychometrist 
for the University of California at San Diego until entering the doctoral program in 
Clinical and Health Psychology at the University of Florida in 1992. She received her 
master of science degree in 1994, following completion of her thesis titled, Pain Severity, 
negative affect, and microstressors as predictors of life interference in TMD patients . Ms. 
Brown is receiving her clinical internship training at McLean Hospital in Belmont, 
Massachusetts. She will graduate with a doctorate in clinical and health psychology with 
an area of concentration in medical psychology. 


I certify that I have read this study and that in my opinion it conforms to 
acceptable standards of scholarly presentation and is fully adequate, in scope and quality, 
as a dissertation for the degree of Doctor of Philosophy. 

lichael E. Robinson, Chair 
Associate Professor of Clinical and Health 

I certify that I have read this study and that in my opinion it conforms to 
acceptable standards of scholarly presentation and is fully adequate, in scope and quality, 
as a dissertation for the degree of Doctor of Philosophy. 

^.^ c# 

Linda/R. Shaw 

Associate Professor of Rehabilitation Counseling 

I certify that I have read this study and that in my opinion it conforms to 
acceptable standards of scholarly presentation and is fully adequate, in scope and quality, 
as a dissertation for the degree of Doctor of Philosophy. 

Cyn&ia D Belar 

Professor of Clinical and Health Psychology 

I certify that I have read this study and that in my opinion it conforms to 
acceptable standards of scholarly presentation and is fully adequate, in scope and quality, 
as a dissertation for the degree of Doctor of Philosophy. 


Henry £f G^eradllion 
Associate Professor of Oral and Maxillofacial 
Surgery and Diagnostic Sciences 

I certify that I have read this study and that in my opinion it conforms to 
acceptable standards of scholarly presentation and is fully^adequate, in scope and quality, 
as a dissertation for the degree of Doctor of Philosophy. 


Duane E. Dede 

Assistant Professor of Clinical and Health 

This dissertation was submitted to the Graduate Faculty of the College of Health 
Professions and to the Graduate School and was accepted as partial fulfillment of the 
requirements for the degree of Doctor of Philosophy. 

May 1998 

l^jt £ WoU^ 

Dean, College of Health Professions 

Dean, Graduate School 


3 1262 08555 3153